In anticipation of the New Year, we summarize the work on the digest and display the trends identified in the process of collecting news on the Python language.

For the year, the news gathering has been brought to the attention and automated to the maximum. Every day, 19 sources are automatically monitored and an average of 10–15 relevant news is gathered from which, the best are subsequently selected and announcements are prepared for digest.
The readers of the digest make a tangible contribution; not a single issue passes without the news they have added.
In the half year that has passed since the weekly digest of news about the python programming language and nearby technologies has ceased to be published in Habré, we have got a permanent audience of readers, made friends with the leading Russian-language python sites, found and learned to handle new sources of information about python in the network, became Constantly publish in the popular public on this topic in contacts, translated and published interesting articles.

Interested in details and figures, welcome under cat.
')

How to keep abreast of all the news.

The main purpose for which the digest was created is to create an aggregator of news and information, both on the python programming language and on branches or modules. During the existence of the digest, approximately 5,235 materials were collected, 1,776 news were translated and published.

Useful information was distributed among the sources as follows:



At the same time, “Different sources” are, for the most part, twitter and what guests sent us via a special form. Social networks (Vkontaktiki, Google+) create noise, but as a source of useful news is almost useless.
The result is that if you don’t want to fall behind the life boiling in the python community, you need to subscribe and read the key of twiiter accounts, read the r / Python top for a week, subscribe to two main mailings, and of course do not ignore Habra. Django line aggregator based on the trends of requests to google on a given topic may be interesting for dzhangists and their sympathizers.

Year of interesting news and trends

During the time spent in searching news, reading articles and participating in public discussions (and this is already more than a year), it was impossible not to single out the news- stars and not catch some trends. Below is a small analysis of the current situation and a selection of interesting articles that we appreciated during the year.

The main trend is seen in the strong development of python in the direction of scientific computing and data analysis. Far from the last place is the remarkable IPython project, which, with the support of powerful computational libraries pandas , numpy , SciKits, allows you to interactively conduct research and conveniently share the results and methods of research / computing. This was noticed in JetBrains - in PyCharm 4 , support and debugging of IPython is improved. Here are a few good articles about this that have been in the digest issues throughout the year:

• Building interactive 3D diagrams in IPython Notebook Based on plotly service
• Using IPython Notebook with Apache Spark. Working through the PySpark module that opens up new scalability limits with Spark and Hadoop technologies.
• Search for bubbles in foam photography. We use a whole set of scientific modules and in addition to the task of searching for bubbles, the direction of foam movement is determined.
• IPython GUI (Jupyter) on the example of image convolution Using a simple html slider to adjust the parameters of the image convolution algorithm in IPython

Faster! Above! Stronger!

Another trend is concern about the performance of calculations and algorithms in general. Here are some interesting approaches. In addition to the above-mentioned modules, in which many numerical algorithms have already been implemented, for example, methods of converting code into machine code are used. Here, the Nuitka project shows itself remarkably and pretends to the ability to assemble any python code into native one by converting it into similar c ++ code and then compiling it. A slightly different approach in the Cython project is that its idea is to compile a subset of the python language into code that can be conveniently used later as a plug-in. Another approach is jit-compiling during execution in a special PyPy interpreter. A version of pypy-stm using the Software Transactional Memory model got to the point where it can actually be used on projects with 2.7. A number of articles about these technologies and their application in practical tasks were really very interesting and even translated to Habré:

• Transferring the execution of a part of highly loaded code from CPython to PyPy It is performed using the multiprocessing module with the path to the new interpreter. Quite an interesting opportunity.
• We write faster code A very curious set of performance tests for different syntax constructs of a language.
• When to choose Jython A small overview and practical example of the advantages of using Jython when integrating with existing Java libraries
• Optimization in python and how they can touch you. The code you write is optimized by the interpreter in various ways during execution. Sometimes this can lead to side effects, some of which are covered in the article and video.
• Why python is slow: look under the hood Review of bottlenecks and their causes in python
• Writing high-performance Cython code
• High-performance extensions for python on C In two articles, the author understands how to speed up the performance of complex calculations by moving them into an extension written in a lower-level language. Here is the second part
• Mathematical optimization problems in Python We understand how to solve problems with many participants, using the technique of stochastic optimization. Essentially, optimization is reduced to finding the best solution to a problem by testing various solutions and comparing them with each other for quality assessment. Usually optimization is used in cases when the number of solutions is too large and it is impossible to sort them all.

2.x vs 3.x

Since 2011, the battle of branches 2.x and 3.x continues. On the one hand, almost all libraries have acceptable support for 3.x, on the other hand, developers are still not in a hurry to switch to the future branch. This is facilitated by the extension of support for the 2.x branch until 2019, as well as backporting of features from the third branch.

• Why Python 3, not Python 2? A question that torments everyone. Answers to this question can be found in this article.
• Python, how Armin Ronher, the author of flask, jinja2, would like to see it, click to share his opinion on the problems in python and in what direction he should have developed.
• PyConRu - Roundtable: Do I need to switch from Python 2 to Python 3
• Why Python 4.0 should not be the same as Python 3.0 Article from the Red Hat developers blog with reflections on

Dive into python

A programming language begins to die as soon as it stops writing deep technical articles and creating training courses. And both of these were abound for the year, with some projects claiming revolutionism.

• CheckIO A very original and fascinating interactive game course on python.
• Learning Python course from learnpython.org Step-by-step learning Python with the ability to execute code directly in the browser
• Special course VMK MSU "Python programming language" For the author of the course (Georgy Kuryachiy), Python turned out to be a real discovery at one time. There was a feeling that Guido planned it according to the principle “it is inconvenient in other languages, it is necessary to make it simpler.” The main focus of this version of the course is just on this simplicity.
• Flask Mega-Tutorial Finally, the cognitive series about web-development on the Flask microfreem is translated
• We teach the old dog new tricks or how I learned to love str.format and refused% Reddit - the channel about python How to love str.format, and give it a chance compared to formatting through%
• Antipattern programming A small number of examples about how not to write code in python
• Cheat Sheet: Writing compatible Python 2-3 code A very useful article on how to write code so that it works equally well in Python 2.6+ and in Python 3.3+
• Top 10 python idioms, which I would like to know earlier A story about interesting, obvious and not very techniques in the python programming language
• And once again about GIL in Python This time regarding the optimization of numerical calculations
• Converting code into beautiful, idiomatic Python Tips on how to give your code beauty.
• Improving package management Overview of several tools (yolk, pip-review, peep ...) that can significantly simplify the life of a developer

Is our web everything?

The shift in focus towards web development, including on mobile platforms, has become a global trend. This is evident in the composition of articles announced on PythonDigest. Most of them are about the web or near it. Here are just a few random articles:

• Django 1.7 The long-awaited release of django 1.7 has been released.
• Why I do not like Flask There is such a popular microframework: Flask. Many people like it: easy and simple to learn, then yes. And the author - absolutely not.
• Build large Flask applications in a realistic environment. Recommendations and basic principles for creating well-supported, scalable and expandable projects.
• Comparison of Django vs Jinja template engines The tests show the impressive superiority of the Jinja2 template engine over the default Django template engine on almost all fronts. Why almost, you ask? And because besides technical metrics there is also the concept of compatibility with already existing applications (admin, cms, reversion ...) and the convenience of expansion.
• Full Django tutorial from beginner to advanced . A tutorial on Django is available for free. It was previously a paid tutorial for everyone. The course will be especially useful for beginners.
• Modern Tornado: distributed image hosting in 30 lines of code Article on the implementation of a simple web server on Tornado

And in conclusion - wishes

This is far from what one would like to tell and could tell, but already the new year is near - it's time to take stock and make a wish. So, since we were good boys and girls all year, we ask Grandfather Frost to separate pieces of the Russian-speaking python community to communicate and communicate even more; for python2 fans to find a compromise with python3 fans; to make kivy the number one platform for mobile apps; to make pythondigest even more friends, help newbies and gurus to join together to develop the community in discussions and in creating new projects and continue to be a consolidating platform and an aggregator of fresh knowledge of Python. Well and peace in the whole world just in case - suddenly it will turn out this time.

Congratulations to all the upcoming New Year!

Many thanks to owlman75 for co-authorship and illustration to the article.